DOCSLIB.ORG

Modeling Network User Behavior: Various Approaches Shidan Xu

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Modeling Network User Behavior: Various Approaches Shidan Xu Computer Science and Artificial Intelligence Laboratory Technical Report MIT-CSAIL-TR-2016-007 June 28, 2016 Modeling Network User Behavior: Various Approaches Shidan Xu massachusetts institute of technology, cambridge, ma 02139 usa — www.csail.mit.edu Modeling User Network Transitions: Various Approaches by Shidan Xu Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Master of Engineering in Electrical Engineering and Computer Science at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2016 c Massachusetts Institute of Technology 2016. All rights reserved. ⃝ Author................................................................ Department of Electrical Engineering and Computer Science May 23, 2016 Certified by............................................................ Professor Karen R. Sollins Principal Research Scientist Thesis Supervisor Accepted by........................................................... Professor Christopher J. Terman Chairman, Masters of Engineering Thesis Committee Modeling Network User Behavior: Various Approaches Shidan Xu May 20, 2016 Abstract This project involves learning to predict users’ mobility within the network topology. Topo- logical mobility, as opposed to physical mobility, can be substantial as a user switches from LTE to wifi network, while moving minimally physically. Our dataset consists of email IMAP logs as they document associated client IP addresses, as well as the clients’ identifiers. Prediction for online mobility is of particular interest to the networks community. If we can predict online mobility with high probability, then new network architecture can be designed to optimize the caching system by minimizing resending packets. We used various approaches and techniques to model the user’s behavior, including probabilistic programming, regression, neural nets, and clustering algorithms. We compare and contrast how models di↵er in their prediction accuracy, speed of convergence, and algorithmic complexity. 1 Acknowledgement I would like to thank Dr. Karen Sollins for giving me an opportunity to work on this MobilityFirst project for my MEng. I loved my time exploring the intricacies in the datasets and I learned so much in the research process. I would like to thank Dr. Karen Sollins and Dr. Steven Bauer for their continued advice and insight into conceiving this project and thesis. I would like to thank Dr. Bauer for his technical help and project focus guidance throughout this project. I’d like to thank Dr. Sollins for all of her help with design decisions and project planning. Without their insights, this project would not have been possible. Their help was invaluable. I would like to thank the MIT Advanced Networks Architecture (ANA) group for having me carry out this research. I would like to thank Tingtao Zhou, Tianfan Xue, and Zhengli Wang for their help with the statistics concepts and machine learning ideas. Lastly, I would like to thank my family for all of their support, love, and help in life. This project was funded in part under NSF Grant 1413973, ”NeTS: Large: Collaborative Re- search: Location-Independent Networks: Evaluation Strategies and Studies”. All code written in this project can be found on Github at All the work can be found on Github at https://github.com/shidanxu/mengvfinal. 2 Contents 1 Introduction 7 1.1 ProjectBackground .................................... 7 1.2 RelationstoMobilityFirst ................................. 7 1.3 Current Internet Has Inefficiency in Topological Distance . 8 1.4 Data Analysis Approaches on Networks Problem . 10 2 Previous Works on Quantitative Network Measurements 11 2.1 Sookhyun Yang’s Three State Markov Model . 11 2.1.1 Choosing the Dataset . 11 2.1.2 Comments on Yang’s Work . 12 2.2 Beverly Applies Machine Learning to Extract Most Important IP Bit for Traffic Con- gestion............................................ 13 2.3 Probabilistic Programming As a Tool For Fast Modeling . 14 3 Reproducing Yang Paper Results 16 3.1 A Peek of the Dataset . 16 3.2 Markov Chain For Transition Prediction . 17 3.3 Evaluate Markov Chain Prediction . 19 3.4 Aspects of Modeling where Markov Chain Falls Short . 22 3.5 Questions that Emerge from Yang Experiment Reproduction . 22 4Modeling 24 4.1 ClusteringFindsThreeGroups .............................. 24 4.2 Clustering Shines Light on Canonical User . 27 4.3 Regression Helps Weighing Factors . 28 4.3.1 Ridge Regression . 29 4.3.2 Logistic Regression . 32 4.4 Neural Net Increases Performance . 35 4.4.1 Background . 35 3 4.4.2 Results and Discussion . 37 4.5 Bayesian Modeling and Probabilistic Programming . 38 4.5.1 Generativevs. DiscriminativeModeling . 38 4.5.2 Probabilistic Programming . 39 5 Approach Comparison 43 5.1 Speed ............................................ 43 5.2 Prediction Accuracy . 44 5.3 Relevance to Dataset . 45 6 Toolkit 47 7 Challenges 49 7.1 Challenges in the Dataset . 49 7.2 Challenges in Technology . 51 8 Contributions and Conclusion 54 8.1 Contributions........................................ 54 8.2 Conclusion ......................................... 55 4 List of Figures 1 Physical proximity does not guarantee topological proximity. Small moves in physical space may be large moves in Network topology. The path from the 4G LTE network to the MIT network goes through US Eastern. 9 2 Sample IMAP entry for 2014.01.28 . 16 3 The distribution of number of transitions from Markov chain and testing set. Sample size 10000 user days each. 20 4 The distribution of number of transitions from training set and testing set. Sample size 10000 user days each. The light blue regions are overlaps of both sets. 21 5 The clustering algorithm output. X-axis is average duration in seconds. Y-axis is session start hour. Average duration length significantly a↵ected the clustering. The greendotsrepresentthecentersofeachcluster. 25 6 X-axis is portion of user entry generated from a wifi network. Y-axis is day of the week. Those were two of the features that did not significantly a↵ected the clustering. 25 7 The K-means clustering algorithm output. The top three factors for separating the dataset are average duration per session, average number of sessions in one day, and the average starting hour for sessions. 26 8 Theaveragenumberofsessionspercluster. 28 9 The long cluster average session length prediction. Each integer on the x-axis is a feature and the y-axis is showing the feature’s corresponding weight. 31 10 The distribution of session durations, in seconds. Note the second peak at the hour mark. ............................................ 32 11 The distribution of session durations, by tags. 34 12 A one hidden-layer feed forward neural net. Image courtesy to Bishop[20]. 36 13 An example PP output using PYMC3. 41 14 Sample session data, each line is a session. The column headers are session start time, sessionendtime,IP,device,userID. 49 15 Exponential distribution implementation in PYMC3. 51 List of Tables 1 Di↵erent kinds of Markov Models[17] . 18 2 Empirical estimates of transition probabilities. Data acquired on training sample (10000 user days). 19 5 3 Empirical estimates of transition probabilities for all users. 27 4 Empirical estimates of transition probabilities for the long duration cluster. 27 5 Comparison of di↵erentnumberofnodesinthehiddenlayer. 38 6 Complexity of di↵erent approaches. 44 7 Prediction Accuracy of di↵erent approaches. 44 8 Comparison of various Python probabilistic programming APIs. 48 6 1 Introduction 1.1 Project Background The Internet is approaching a historic inflection point, with mobile platforms and applications re- placing the fixed-host/server model that has dominated the Internet since its inception [14]. This gradual shift produces an opportunity to design a new network focused on mobile users. The prob- lem this thesis investigates is motivated by the MobilityFirst project in the FIND (Future Internet Design) initiative [7], which asks the question of what the requirements should be for a global net- work 15 years from now, and how we can build such a network if we are not constrained by the current Internet. 1.2 Relations to MobilityFirst The MobilityFirst projects presents several major design goals [14]: • Dynamic hosting to provide scalable network mobility. • The robustness of the wireless transfer medium. • Reinforce network security and privacy for both mobile and wired networks. • Provide context-aware mobile services. Mobile users often switch between di↵erent networks (such as cellular and wifi) as they move geographically. Traditional networks such as the Internet were designed more for a static user. In designing a new network that is more suited to the mobile users, we need to evaluate this switching behavior. In this thesis project, we evaluate the user’s network transition behavior, in order to understand which users should the new network address. This relates to the first design goal. By 7 providing a clearer understanding of the user behavior, new networks can be designed to configure hosts that is more suitable for the user. One aspect of suitable is to minimize the di↵erence in topological distance and geographical distance. We expand this point next. 1.3 Current Internet Has Inefficiency in Topological Distance Developers of the Internet made some design choices with assumptions. One particular assumption is that network mobility is similar to mobility in geography[3]. However, in real world cases, this assumption is often violated. For example,
Load more

Recommended publications
  • University of North Carolina at Charlotte Belk College of Business
    University of North Carolina at Charlotte Belk College of Business BPHD 8220 Financial Bayesian Analysis Fall 2019 Course Time: Tuesday 12:20 - 3:05 pm Location: Friday Building 207 Professor: Dr. Yufeng Han Office Location: Friday Building 340A Telephone: (704) 687-8773 E-mail: [email protected] Office Hours: by appointment Textbook: Introduction to Bayesian Econometrics, 2nd Edition, by Edward Greenberg (required) An Introduction to Bayesian Inference in Econometrics, 1st Edition, by Arnold Zellner Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan, 2nd Edition, by Joseph M. Hilbe, de Souza, Rafael S., Emille E. O. Ishida Bayesian Analysis with Python: Introduction to statistical modeling and probabilistic programming using PyMC3 and ArviZ, 2nd Edition, by Osvaldo Martin The Oxford Handbook of Bayesian Econometrics, 1st Edition, by John Geweke Topics: 1. Principles of Bayesian Analysis 2. Simulation 3. Linear Regression Models 4. Multivariate Regression Models 5. Time-Series Models 6. State-Space Models 7. Volatility Models Page | 1 8. Endogeneity Models Software: 1. Stan 2. Edward 3. JAGS 4. BUGS & MultiBUGS 5. Python Modules: PyMC & ArviZ 6. R, STATA, SAS, Matlab, etc. Course Assessment: Homework assignments, paper presentation, and replication (or project). Grading: Homework: 30% Paper presentation: 40% Replication (project): 30% Selected Papers: Vasicek, O. A. (1973), A Note on Using Cross-Sectional Information in Bayesian Estimation of Security Betas. Journal of Finance, 28: 1233-1239. doi:10.1111/j.1540- 6261.1973.tb01452.x Shanken, J. (1987). A Bayesian approach to testing portfolio efficiency. Journal of Financial Economics, 19(2), 195–215. https://doi.org/10.1016/0304-405X(87)90002-X Guofu, C., & Harvey, R.
    [Show full text]
  • An Introduction to Data Analysis Using the Pymc3 Probabilistic Programming Framework: a Case Study with Gaussian Mixture Modeling
    An introduction to data analysis using the PyMC3 probabilistic programming framework: A case study with Gaussian Mixture Modeling Shi Xian Liewa, Mohsen Afrasiabia, and Joseph L. Austerweila aDepartment of Psychology, University of Wisconsin-Madison, Madison, WI, USA August 28, 2019 Author Note Correspondence concerning this article should be addressed to: Shi Xian Liew, 1202 West Johnson Street, Madison, WI 53706. E-mail: [email protected]. This work was funded by a Vilas Life Cycle Professorship and the VCRGE at University of Wisconsin-Madison with funding from the WARF 1 RUNNING HEAD: Introduction to PyMC3 with Gaussian Mixture Models 2 Abstract Recent developments in modern probabilistic programming have offered users many practical tools of Bayesian data analysis. However, the adoption of such techniques by the general psychology community is still fairly limited. This tutorial aims to provide non-technicians with an accessible guide to PyMC3, a robust probabilistic programming language that allows for straightforward Bayesian data analysis. We focus on a series of increasingly complex Gaussian mixture models – building up from fitting basic univariate models to more complex multivariate models fit to real-world data. We also explore how PyMC3 can be configured to obtain significant increases in computational speed by taking advantage of a machine’s GPU, in addition to the conditions under which such acceleration can be expected. All example analyses are detailed with step-by-step instructions and corresponding Python code. Keywords: probabilistic programming; Bayesian data analysis; Markov chain Monte Carlo; computational modeling; Gaussian mixture modeling RUNNING HEAD: Introduction to PyMC3 with Gaussian Mixture Models 3 1 Introduction Over the last decade, there has been a shift in the norms of what counts as rigorous research in psychological science.
    [Show full text]
  • Is Structure Based Drug Design Ready for Selectivity Optimization?
    bioRxiv preprint doi: https://doi.org/10.1101/2020.07.02.185132; this version posted July 3, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. PrEPRINT AHEAD OF SUBMISSION — July 2, 2020 1 IS STRUCTURE BASED DRUG DESIGN READY FOR 2 SELECTIVITY optimization? 1,2,† 2 3 4 4 3 SteVEN K. Albanese , John D. ChoderA , AndrEA VOLKAMER , Simon Keng , Robert Abel , 4* 4 Lingle WANG Louis V. Gerstner, Jr. GrADUATE School OF Biomedical Sciences, Memorial Sloan Kettering Cancer 5 1 Center, NeW York, NY 10065; Computational AND Systems Biology PrOGRam, Sloan Kettering 6 2 Institute, Memorial Sloan Kettering Cancer Center, NeW York, NY 10065; Charité – 7 3 Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin; Schrödinger, NeW York, NY 10036 8 4 [email protected] (LW) 9 *For CORRespondence: †Schrödinger, NeW York, NY 10036 10 PrESENT ADDRess: 11 12 AbstrACT 13 Alchemical FREE ENERGY CALCULATIONS ARE NOW WIDELY USED TO DRIVE OR MAINTAIN POTENCY IN SMALL MOLECULE LEAD 14 OPTIMIZATION WITH A ROUGHLY 1 Kcal/mol ACCURACY. Despite this, THE POTENTIAL TO USE FREE ENERGY CALCULATIONS 15 TO DRIVE OPTIMIZATION OF COMPOUND SELECTIVITY AMONG TWO SIMILAR TARGETS HAS BEEN RELATIVELY UNEXPLORED IN 16 PUBLISHED studies. IN THE MOST OPTIMISTIC scenario, THE SIMILARITY OF BINDING SITES MIGHT LEAD TO A FORTUITOUS 17 CANCELLATION OF ERRORS AND ALLOW SELECTIVITY TO BE PREDICTED MORE ACCURATELY THAN AffiNITY. Here, WE ASSESS 18 THE ACCURACY WITH WHICH SELECTIVITY CAN BE PREDICTED IN THE CONTEXT OF SMALL MOLECULE KINASE inhibitors, 19 CONSIDERING THE VERY SIMILAR BINDING SITES OF HUMAN KINASES CDK2 AND CDK9, AS WELL AS ANOTHER SERIES OF 20 LIGANDS ATTEMPTING TO ACHIEVE SELECTIVITY BETWEEN THE MORE DISTANTLY RELATED KINASES CDK2 AND ERK2.
    [Show full text]
  • Migrating MATLAB®To Python
    Migrating MATLAB® to Python Strategies, Comparisons and a Guide to Converting for Experts Migrating MATLAB® to Python Strategies, Comparisons and a Guide to Converting for Experts Alexandre Chabot-Leclerc Enthought, Inc. ©2020 Enthought, Inc. Written by Enthought, Inc. All Rights Reserved. Use only permitted under license. Copying, sharing, redistributing, or other unauthorized use strictly prohibited. All trademarks and registered trademarks are the property of their respective owners. MATLAB and Simulink are registered trademark of The MathWorks, Inc. Enthought, Inc. 200 W Cesar Chavez St Suite 202 Austin, TX 78701 United States www.enthought.com Version 1.2.0 Migrating MATLAB® to Python B Introduction 1 Why Python 2 Diff erences Between Python and MATLAB® 4 Fundamental Data Types 4 Organizing Code in Packages, not Toolboxes 6 Syntax 6 Indexing and Slicing: Why Zero-Based Indexing 8 NumPy Arrays Are Not Matrices 10 Programming Paradigm: Object-Oriented vs. Procedural 13 Anti-Patterns 15 How Do I? 17 Load Data 17 Signal Processing 19 Linear Algebra 19 Machine Learning 20 Statistical Analysis 21 Image Processing and Computer Vision 21 Optimization 22 Natural Language Processing 22 Data Visualization 22 Save Data 25 What Else? 26 Strategies for Converting to Python 27 From the Bottom Up: Converting One Function at a Time 27 From the Top Down: Calling Python from MATLAB® 33 What Next? 37 Acknowledgments 37 Appendix 37 Code Example: Profiling Contiguous Array Operations 37 Complete Version of main.py, in Chapter 4 38 References 39 Accelerate your Python migration with Enthought’s Python for Scientists and Engineers training course! Migrating MATLAB® to Python C Introduction This document will guide you through the transition from MATLAB® to Python.
    [Show full text]
  • Probabilistic Programming in Python Using Pymc3
    Probabilistic programming in Python using PyMC3 John Salvatier1, Thomas V. Wiecki2 and Christopher Fonnesbeck3 1 AI Impacts, Berkeley, CA, United States 2 Quantopian Inc, Boston, MA, United States 3 Department of Biostatistics, Vanderbilt University, Nashville, TN, United States ABSTRACT Probabilistic programming allows for automatic Bayesian inference on user-defined probabilistic models. Recent advances in Markov chain Monte Carlo (MCMC) sampling allow inference on increasingly complex models. This class of MCMC, known as Hamiltonian Monte Carlo, requires gradient information which is often not readily available. PyMC3 is a new open source probabilistic programming framework written in Python that uses Theano to compute gradients via automatic differentiation as well as compile probabilistic programs on-the-fly to C for increased speed. Contrary to other probabilistic programming languages, PyMC3 allows model specification directly in Python code. The lack of a domain specific language allows for great flexibility and direct interaction with the model. This paper is a tutorial-style introduction to this software package. Subjects Data Mining and Machine Learning, Data Science, Scientific Computing and Simulation Keywords Bayesian statistic, Probabilistic Programming, Python, Markov chain Monte Carlo, Statistical modeling INTRODUCTION Probabilistic programming (PP) allows for flexible specification and fitting of Bayesian Submitted 9 September 2015 statistical models. PyMC3 is a new, open-source PP framework with an intuitive and Accepted 8 March 2016 readable, yet powerful, syntax that is close to the natural syntax statisticians use to Published 6 April 2016 describe models. It features next-generation Markov chain Monte Carlo (MCMC) Corresponding author Thomas V. Wiecki, sampling algorithms such as the No-U-Turn Sampler (NUTS) (Hoffman & Gelman, [email protected] 2014), a self-tuning variant of Hamiltonian Monte Carlo (HMC) (Duane et al., 1987).
    [Show full text]
  • Modified Hamiltonian Monte Carlo for Bayesian Inference
    Modified Hamiltonian Monte Carlo for Bayesian Inference Tijana Radivojević1;2;3, Elena Akhmatskaya1;4 1BCAM - Basque Center for Applied Mathematics 2Biological Systems and Engineering Division, Lawrence Berkeley National Laboratory 3DOE Agile Biofoundry 4IKERBASQUE, Basque Foundation for Science July 26, 2019 Abstract The Hamiltonian Monte Carlo (HMC) method has been recognized as a powerful sampling tool in computational statistics. We show that performance of HMC can be significantly improved by incor- porating importance sampling and an irreversible part of the dynamics into a chain. This is achieved by replacing Hamiltonians in the Metropolis test with modified Hamiltonians, and a complete momen- tum update with a partial momentum refreshment. We call the resulting generalized HMC importance sampler—Mix & Match Hamiltonian Monte Carlo (MMHMC). The method is irreversible by construc- tion and further benefits from (i) the efficient algorithms for computation of modified Hamiltonians; (ii) the implicit momentum update procedure and (iii) the multi-stage splitting integrators specially derived for the methods sampling with modified Hamiltonians. MMHMC has been implemented, tested on the popular statistical models and compared in sampling efficiency with HMC, Riemann Manifold Hamiltonian Monte Carlo, Generalized Hybrid Monte Carlo, Generalized Shadow Hybrid Monte Carlo, Metropolis Adjusted Langevin Algorithm and Random Walk Metropolis-Hastings. To make a fair com- parison, we propose a metric that accounts for correlations among samples and weights,
    [Show full text]
  • Probabilistic Programming
    Probabilistic programming Probabilistic programming (PP) is a programming paradigm in which probabilistic models are specified and inference for these models is performed automatically.[1] It represents an attempt to unify probabilistic modeling and traditional general purpose programming in order to make the former easier and more widely applicable.[2][3] It can be used to create systems that help make decisions in the face of uncertainty. Programming languages used for probabilistic programming are referred to as "probabilistic programming languages" (PPLs). Contents Applications Probabilistic programming languages Relational List of probabilistic programming languages Difficulty See also Notes External links Applications Probabilistic reasoning has been used for a wide variety of tasks such as predicting stock prices, recommending movies, diagnosing computers, detecting cyber intrusions and image detection.[4] However, until recently (partially due to limited computing power), probabilistic programming was limited in scope, and most inference algorithms had to be written manually for each task. Nevertheless, in 2015, a 50-line probabilistic computer vision program was used to generate 3D models of human faces based on 2D images of those faces. The program used inverse graphics as the basis of its inference method, and was built using the Picture package in Julia.[4] This made possible "in 50 lines of code what used to take thousands".[5][6] The Gen probabilistic programming library (also written in Julia) has been applied to vision and robotics tasks.[7] More recently, the probabilistic programming systems Turing.jl has been applied in various pharmaceutical and economics applications.[8] Probabilistic programming in Julia has also been combined with differentiable programming by combining the Julia package Zygote.jl with Turing.jl.
    [Show full text]
  • Greta: Simple and Scalable Statistical Modelling in R
    greta: simple and scalable statistical modelling in R Nick Golding1 1 School of BioSciences, University of Melbourne DOI: 10.21105/joss.01601 Software • Review Summary • Repository • Archive Statistical modelling is used throughout the sciences. Often, statistical analyses require cus- Submitted: 26 July 2019 tom models that cannot be fitted using off-the shelf statistical software. These models can Published: 12 August 2019 be specified in a statistical syntax and can then be automatically fit to data usingmethods License such as Markov Chain monte Carlo (MCMC) and maximum likelihood. This lets users fo- Authors of papers retain cus on the statistical nature of the model, rather than implementation details and inference copyright and release the work procedures. Since the development of the widely successful WinBUGS (later developed as under a Creative Commons OpenBUGS; Spiegelhalter, Thomas, Best, & Lunn (2014)) a number of alternative software Attribution 4.0 International packages for custom statistical modelling have been introduced, including JAGS, Stan, and License (CC-BY). NIMBLE (Carpenter et al., 2017; de Valpine et al., 2017; Plummer & others, 2003). In these software packages, users typically write out models in a domain-specific language, which is then compiled into computational code. Though see the Python packages PyMC and Edward (Salvatier, Wiecki, & Fonnesbeck, 2016; Tran et al., 2016) in which models are specified in Python code. With increasing quantities of data, complexity, and realism of statistical models that users wish to build with these software, there is a push for software that scales better with data size and model complexity. More recently, custom statistical modelling software has focussed on methods such as Hamiltonian Monte Carlo (rather than Gibbs samplers) in order to improve to computational efficiency.
    [Show full text]
  • Hands-On Bayesian Neural Networks - a Tutorial for Deep Learning Users
    Hands-on Bayesian Neural Networks - a Tutorial for Deep Learning Users LAURENT VALENTIN JOSPIN, University of Western Australia WRAY BUNTINE, Monash University FARID BOUSSAID, University of Western Australia HAMID LAGA, Murdoch university MOHAMMED BENNAMOUN, University of Western Australia Modern deep learning methods have equipped researchers and engineers with incredibly powerful tools to tackle problems that previously seemed impossible. However, since deep learning methods operate as black boxes, the uncertainty associated with their predictions is often challenging to quantify. Bayesian statistics offer a formalism to understand and quantify the uncertainty associated with deep neural networks predictions. This paper provides a tutorial for researchers and scientists who are using machine learning, especially deep learning, with an overview of the relevant literature and a complete toolset to design, implement, train, use and evaluate Bayesian neural networks. CCS Concepts: • Mathematics of computing ! Probability and statistics; • Computing methodologies ! Neural networks; Bayesian network models; Ensemble methods; Regularization. Additional Key Words and Phrases: Bayesian methods, Bayesian Deep Learning, Approximate Bayesian methods ACM Reference Format: Laurent Valentin Jospin, Wray Buntine, Farid Boussaid, Hamid Laga, and Mohammed Bennamoun. 2020. Hands-on Bayesian Neural Networks - a Tutorial for Deep Learning Users. ACM Comput. Surv. 1, 1 (July 2020), 35 pages. https://doi.org/10.1145/nnnnnnn.nnnnnnn 1 INTRODUCTION Deep learning has led to a revolution in machine learning, providing solutions to tackle more and more complex and challenging real-life problems. However, deep learning models are prone to overfitting, which adversely affects their generalization capabilities. Deep learning models also tend to be overconfident about their predictions (when they do provide a confidence interval).
    [Show full text]
  • Automatic Differentiation in Machine Learning: a Survey
    Automatic Differentiation in Machine Learning: a Survey Atılım G¨une¸sBaydin [email protected] Department of Engineering Science University of Oxford Oxford OX1 3PJ, United Kingdom Barak A. Pearlmutter [email protected] Department of Computer Science National University of Ireland Maynooth Maynooth, Co. Kildare, Ireland Alexey Andreyevich Radul [email protected] Department of Brain and Cognitive Sciences Massachusetts Institute of Technology Cambridge, MA 02139, United States Jeffrey Mark Siskind [email protected] School of Electrical and Computer Engineering Purdue University West Lafayette, IN 47907, United States Abstract Derivatives, mostly in the form of gradients and Hessians, are ubiquitous in machine learn- ing. Automatic differentiation (AD), also called algorithmic differentiation or simply “auto- diff”, is a family of techniques similar to but more general than backpropagation for effi- ciently and accurately evaluating derivatives of numeric functions expressed as computer programs. AD is a small but established field with applications in areas including compu- arXiv:1502.05767v4 [cs.SC] 5 Feb 2018 tational fluid dynamics, atmospheric sciences, and engineering design optimization. Until very recently, the fields of machine learning and AD have largely been unaware of each other and, in some cases, have independently discovered each other’s results. Despite its relevance, general-purpose AD has been missing from the machine learning toolbox, a situ- ation slowly changing with its ongoing adoption under the names “dynamic computational graphs” and “differentiable programming”. We survey the intersection of AD and machine learning, cover applications where AD has direct relevance, and address the main imple- mentation techniques. By precisely defining the main differentiation techniques and their interrelationships, we aim to bring clarity to the usage of the terms “autodiff”, “automatic differentiation”, and “symbolic differentiation” as these are encountered more and more in machine learning settings.
    [Show full text]
  • Tutorial on Probabilistic Programming with Pymc3
    185.A83 Machine Learning for Health Informatics 2017S, VU, 2.0 h, 3.0 ECTS Tutorial 02 - 04.04.2017 Tutorial on Probabilistic Programming with PyMC3 [email protected] http://hci-kdd.org/machine-learning-for-health-informatics-course Holzinger Group hci-kdd.org 1 MAKE Health T01 Schedule ▪ 01. Introduction to Probabilistic Programming ▪ 02. PyMC3 ▪ 03. linear regression – the Bayesian way ▪ 04. generalized linear models with PyMC3 Holzinger Group hci-kdd.org 2 MAKE Health T01 01. Introduction and Overview ▪ Probabilistic Programming (PP) ▪ allows automatic Bayesian inference ▪ on complex, user-defined probabilistic models ▪ utilizing “Markov chain Monte Carlo” (MCMC) sampling ▪ PyMC3 ▪ a PP framework ▪ compiles probabilistic programs on-the-fly to C ▪ allows model specification in Python code Salvatier J, Wiecki TV, Fonnesbeck C. (2016) Probabilistic programming in Python using PyMC3. PeerJ Computer Science 2:e55 https://doi.org/10.7717/peerj-cs.55 Holzinger Group hci-kdd.org 3 MAKE Health T01 Properties of Probabilistic Programs ▪ IS NOT ▪ Software that behaves probabilistically ▪ General programming language ▪ IS ▪ Toolset for statistical / Bayesian modeling ▪ Framework to describe probabilistic models ▪ Tool to perform (automatic) inference ▪ Closely related to graphical models and Bayesian networks ▪ Extension to basic language (e.g. PyMC3 for Python) “does in 50 lines of code what used to take thousands” Kulkarni, T. D., Kohli, P., Tenenbaum, J. B. & Mansinghka, V. Picture: A probabilistic programming language for scene perception. in Proceedings of the ieee conference on computer vision and pattern recognition 4390–4399 (2015). Holzinger Group hci-kdd.org 4 MAKE Health T01 Probabilistic Programs ▪ Machine learning algorithms / models often a black box PP “open box” ▪ Simple approach 1.
    [Show full text]
  • Enhancing Sampling in Computational Statistics [0.2Cm]
    Tesis Doctoral / Doktorego Tesia Enhancing Sampling in Computational Statistics Using Modified Hamiltonians Autora / Egilea: Tijana Radivojevi¢ Directores / Zuzendariek: Prof. Elena Akhmatskaya Prof. Enrico Scalas 2016 Doctoral Thesis Enhancing Sampling in Computational Statistics Using Modified Hamiltonians Author: Tijana Radivojevi¢ Supervisors: Prof. Elena Akhmatskaya Prof. Enrico Scalas 2016 This research was carried out at the Basque Center for Applied Mathematics (BCAM) within the Group Modelling and Simulation in Life and Materials Sciences. The research was supported by the Spanish Ministry of Education, Culture and Sport (MECD) within the FPU-2012 program (Formación del Profesorado Universitario) under Grant FPU12/05209 and also by the Basque Government through the BERC 2014-2017 program and by the Spanish Ministry of Economy and Competitiveness MINECO: BCAM Severo Ochoa excellence accreditation SEV-2013-0323 and Grant Retosenintegraciónnumérica: delasestructurasalgebráicasasimulacionesMontecarlo MTM2013– 46553–C3–1–P. ii Abstract The Hamiltonian Monte Carlo (HMC) method has been recognized as a powerful sampling tool in computational statistics. In this thesis, we show that performance of HMC can be dramati- cally improved by replacing Hamiltonians in the Metropolis test with modified Hamiltonians, and a complete momentum update with a partial momentum refreshment. The resulting general- ized HMC importance sampler, which we called Mix & Match Hamiltonian Monte Carlo (MMHMC), arose as an extension of the Generalized Shadow Hybrid Monte Carlo (GSHMC) method, previ- ously proposed for molecular simulation. The MMHMC method adapts GSHMC specifically to computational statistics and enriches it with new essential features: (i) the eicient algorithms for computation of modified Hamiltonians; (ii) the implicit momentum update procedure and (iii) the two-stage splitting integration schemes specially derived for the methods sampling with modified Hamiltonians.
    [Show full text]